Title :
An improved method for feature weighting to document categorization
Author :
Pei, Zhili ; Zhou, Yuxin ; Wang, Lihua ; Liu, Lisha ; Lu, Yinan ; Kong, Ying
Author_Institution :
Coll. of Comput. Sci. & Technol., Inner Mongolia Univ. for the Nationalites, Tongliao, China
Abstract :
Based on the significance of feature frequency, an improved weighting method for feature frequency is proposed, which considers the decision-making information in evaluating the contribution of feature frequency to document categorization. Experimental results show that this method could improve the space distribution of the document set. Furthermore, the values of macro accuracy, macro recall rate and macroF1 are all improved significantly.
Keywords :
decision making; document handling; decision-making information; document categorization; feature frequency; feature weighting; Automation; Computer science; Decision making; Educational institutions; Frequency; Fuzzy sets; Machine learning; Mechanical engineering; Set theory; Workstations; document categorization; feature weighting; real rough set; tf.idf;
Conference_Titel :
Future Computer and Communication (ICFCC), 2010 2nd International Conference on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4244-5821-9
DOI :
10.1109/ICFCC.2010.5497607
